Topline · 12 Apr 2026 · From the week of 6 April
How Top VCs Pick Winners In 2026 | Cassie Young, General Partner @ Primary Venture Partners
These are notes on the conversation, checked against its transcript. The episode itself has the full discussion.
In brief
Topline hosts AJ Bruno and Asad Zaman, with Sam Jacobs, interview Cassie Young, a General Partner at Primary Venture Partners, about the firm's $625 million seed fund and its view that seed funds can stay large by adding check writers. The conversation covers AI-native CRM competition with Salesforce, the size of the white-collar labor market as a target, and a possible gross retention reckoning for AI-native companies. Cassie argues that a strong go-to-market CEO is not enough without technologists, and that implementation and customer success are becoming product problems.
For founders
- A zero CAC CEO is not enough on its own, so Primary asks early who the technical co-founder is and what they will do for the business.
- Cassie says the capital needed to win a competitive seed deal is 20 to 25% higher than it was 18 months ago.
- Cassie expects AI-native businesses' lower gross retention to produce a reckoning in year one or two, though the timing is uncertain.
- Cassie tests market demand by reaching out to six to 10 strong leaders and watching how quickly they reply and whether they take a demo.
- Cassie says the best organizations treat forward-deployed work as a source of product learning and productize implementation quickly, and some portfolio companies are moving implementation into the product org.
For revenue leaders
- Cassie recommends time to expansion as a leading indicator, because enterprise renewals can take about a year.
- Cassie says gross and net retention should be reported separately for current ICP customers and legacy customers, since early churn outside the ICP can be healthy.
- Cassie says proactive customer success needs products that spot configuration problems before customers raise them.
- AJ Bruno says his company took about three years to move its revenue mix from 70% SMB to 70% mid-market.
- Cassie says moving from a revenue focus to a customer count focus is a mindset shift some companies are making as they focus on retention.
What was said 25, most useful first
The capital needed to win a seed deal is 20 to 25% higher than it was 18 months ago, even before counting mega seed rounds. Listen
Cassie says the quantum of capital required to win a seed deal is 20 to 25% higher than it was 18 months ago. She says she is talking about ordinary competitive seed deals, not the crazy mega seed rounds.
“the quantum of capital required to win a seed deal now vis-à-vis where it was 18 months ago is 20 to 25% higher”
Listen to the episode Fundraising & investors Link to this Report a problem
A zero CAC CEO is not enough on its own, so Primary asks who the technical co-founder is. Listen
Cassie says that for a while she may have over-rotated on zero CAC CEOs in a few instances, meaning founders who are well networked in their category with a clear right to win. She says without technologists who can keep the company a step ahead, such a CEO may get in the door but will not keep the share they win early. Primary's follow-up question is who the technical co-founder is and what they will do for the business.
“It may get you in the door, but to your point, it's not going to ensure that you kind of keep, you know, the share that you get out of the gate.”
Listen to the episode Hiring & team building Link to this Report a problem
AI-native businesses showed lower gross retention than legacy SaaS companies in one dataset of earlier-stage companies. Listen
Cassie says Kyle Poyar's data, which focused on companies earlier in their journeys, validated her hypothesis that AI-native businesses have lower gross retention than legacy SaaS counterparts. She had hoped to have hard data when she wrote her gross retention piece, but Kyle had none at the time and published his analysis later. The figures themselves were not given in the conversation.
“his data validated my hypothesis which is that the AI native businesses had lower gross retention than the legacy SAS counterparts”
Listen to the episode Metrics & finance Link to this Report a problem
Time to expansion is a leading indicator of retention when enterprise renewals take about a year. Listen
Cassie says enterprise renewals take about a year, so Primary's portfolio companies look at other leading indicators. One is time to expansion, which can mean time to commercial expansion or time to use-case expansion. She says she pushes portfolio founders on this because retention cannot be read from renewals alone early on.
“it's going to take you a year to get a renewal in an enterprise environment like what are the other leading indicators that you can look to and so one of the things that we talk about is just the time to expansion”
Listen to the episode Retention & customer success Link to this Report a problem
Board reporting should split gross and net retention between current ICP customers and legacy customers outside the ICP. Listen
Cassie says by Series B-plus board meetings, companies often break out gross and net retention by current ICP customers versus legacy ICP customers. Early churn can be a normal and healthy part of finding the right customer, but companies should be honest about which lost customers were inside the ICP. She warns against losing customers who should be inside it.
“in series B plus board meetings you see the numbers for both GRR and NR broken out by current ICP customers versus legacy ICP customers”
Listen to the episode Metrics & finance Link to this Report a problem
Cassie tests market demand by messaging six to 10 strong leaders and watching how fast they reply and whether they take a demo. Listen
Cassie says one of her best signals is market feedback: how quickly six to 10 great leaders respond to her outreach and whether they will take a demo, which shows how hair on fire the problem is. For one company she was seven for seven on demos, but the feedback convinced her buyers would use it for year one and then build it themselves, so she passed.
“I always say like one of my best signals is just the the market feedback where if I reach out to six to 10 great leaders how quickly did they get back to me and their willingness to take a demo”
Listen to the episode Pipeline & demand generation Link to this Report a problem
Moving revenue from 70% SMB to 70% mid-market took about three years. Listen
AJ Bruno says it took his company about three years to move from a revenue mix where SMB was 70% to one where mid-market is now 70%. He says he does not yet know how AI will accelerate that shift.
“it's taken us honestly three years to move from SMB where 70% was of our revenue now today 70% is in the midmarket”
Listen to the episode Strategy & market Link to this Report a problem
Forward-deployed engineers only form a true model when implementation work feeds a feedback loop into product. Listen
A host says many AI companies have forward-deployed engineers but no plumbing connecting them to product, so the work is just implementation. He says the true FDE model is one where that work feeds back into what gets built, which is what made the Palantir model special. Cassie agrees, saying Primary has discussed the Palantir model internally and that Palantir built product off it rather than just selling services.
“The whole FTE concept was it was going to help Palantir figure out what to build and that feedback loop is what was so special about the whole model”
About 2.5 to 3% of institutionally backed seed deals ultimately become unicorns, which Cassie uses to judge how much room a firm has to hunt. Listen
Cassie describes the seed-to-unicorn ratio as historically 2.5 to 3 percent. She frames it as a total addressable market question for investors: how much room they have to hunt for deals and how hard they must then find the great ones. It describes how investors size their opportunity set.
“2 and a half to 3% of all deals that have been institutionally backed at seed will ultimately become unicorn companies”
Listen to the episode Fundraising & investors Link to this Report a problem
Primary asks whether it would bet against a founder when a deal reaches the finish line. Listen
Cassie says her partner Ben uses the line 'would you bet against this person?' and that when a deal reaches the finish line, the firm treats it as a reflection exercise. She presents it as a fundamental question the firm asks itself at that point. It is investor diligence shorthand.
“would you bet against this person? And at the end of the day, like when we're at the finish line with a deal, that's a fundamental question that we're asking ourselves”
Listen to the episode Fundraising & investors Link to this Report a problem
Cassie expects an AI retention reckoning in years one or two, after which good companies will improve, though the timing is uncertain. Listen
Cassie says if she were a betting woman she suspects companies with weak retention will face a reckoning in year one or two, and then the good ones will get better, so cohorts should improve over time. She says it might not be next month or next year, but it is ultimately going to come. She also says the enterprise buyer may be slower to consolidate tools than she expected.
“if I were a betting woman I suspect that for companies where that happens It'll be a reckoning in year one or two and then the good companies will they'll wise up about it”
Listen to the episode Retention & customer success Link to this Report a problem
Some companies are moving from a revenue focus to a customer count focus as they lean into retention. Listen
Cassie describes an executive who previously talked about revenue constantly and now, at One Mind, is focused on customer count, asking how to make customers happy and grow them. Cassie calls this an important mentality shift as people lean into retention, and says she still worries about the retention risk.
“with one mind she's obsessed with customer count right and like that customer how do you make them happy right how do you grow them”
Listen to the episode Retention & customer success Link to this Report a problem
Cassie expected CIOs to consolidate AI tool sprawl in 2026, but enterprises have been slower to do it. Listen
Cassie says she expected individual contributors and managers to have budgets to experiment with AI tools, and that CIOs would come in during 2026 to consolidate under a few vendors, which she called a clearing house for gross retention. She says this has not happened as quickly as she thought, because some large enterprises are so eager for employees to use AI that consolidation is not on their minds.
“I do still think that's going to happen it hasn't happened as quickly as I would have thought”
Listen to the episode Strategy & market Link to this Report a problem
The white-collar labor pool is about $6 trillion, roughly 20 times what is spent on enterprise software. Listen
Cassie cites the white-collar labor pool as $6 trillion, 20 times enterprise software spend, and says that even if the full 20x is not realized the magnitude is still material. She also cites software at 8.5% of the S&P and 1% of GDP to argue there is room beyond the SaaS market. These are figures she pulled for Primary's annual LP meeting.
“the white collar labor pool is $6 trillion. That's 20 times what's been spent on enterprise software”
Listen to the episode Strategy & market Link to this Report a problem
Intercom's deliberate move to cannibalize its software revenue to go all in on AI is an exception rather than the rule. Listen
Cassie praises Intercom for deliberately cannibalizing its software revenue to go all in on AI, calling it an amazing story. She says she expects Intercom to be the exception, not the rule. A host adds that Intercom's story is not yet written.
“I think Intercom's probably going to be the exception, not the rule.”
Listen to the episode Strategy & market Link to this Report a problem
AI-native CRM upstarts have not yet solved the move from mid-market to up-market. Listen
Cassie says that in diligence on AI-native CRM she heard that the upstarts are so native they have not nailed moving from mid-market to up-market. She says they have to believe someone will work out how, and that these companies must do the unscalable things to make it happen. Many are going after Y Combinator and Series B companies.
“what I heard was like all of the upstarts are so native that like they haven't nailed this mid-market to up market yet”
Listen to the episode Strategy & market Link to this Report a problem
An AI-native CRM needs a compound startup approach to take share; Cassie sees a wedge-first path as harder. Listen
Cassie contrasts a classical wedge play, where a company takes one clear wedge with a broad platform vision, with a compound startup approach that spans the whole customer lifecycle. She says you have to take the compound approach to take share, and that a sales-coaching company building an invisible CRM is a harder path, which is her personal conviction.
“A native CRM like you kind of have to go at it from the compound startup approach if you're really going to take share”
Listen to the episode Strategy & market Link to this Report a problem
Partial-platform AI tools now look more like features than companies, because full platforms make them unnecessary. Listen
Cassie says full-platform plays are more compelling than partial-platform plays, which now feel more like features, because an end-to-end platform does the work that point tools did. Separately, she says buyers like having a throat to choke for security and governance, which makes her less worried about DIY replacing platforms. She says eight months earlier she would probably have backed a company solving one part of the problem, but now she expects buyers to adopt it for year one and then build it themselves.
“in these full platform plays versus things that even looked like partial platform plays now feel more like features”
Listen to the episode Strategy & market Link to this Report a problem
When assessing compound startups, Cassie looks at how the company manages design partners, how fast it learns and how quickly its engineering roadmap moves. Listen
Cassie says Primary looks closely at design partners at seed for compound startups, asking how the founders manage them, how they learn and what their pace is. She also looks at the velocity of the engineering roadmap and whether the company launches things with enough breadth and depth to show it is learning from customers.
“How are you managing those? How are you learning? What is your pace?”
A CRM's two real modes are data and distribution, and new AI-native upstarts have neither. Listen
AJ Bruno says the CRM has two big modes, data and distribution, and that the product itself is becoming commoditized. He says Salesforce has locked down its data, and that distribution is both the challenge and the opportunity for startups trying to move from SMB into mid-market.
“The CRM has two real big modes, data and distribution. And no no upstarts or seeds have have those.”
Listen to the episode Strategy & market Link to this Report a problem
No product advantage lasts anymore, so staying one step ahead means linking go-to-market signals to a lean product team. Listen
Asad Zaman describes a founder with about $30M NR who merged engineering, product and design into one unit working closely with go-to-market. The go-to-market team listens to the market and feeds the product team, which can build and ship quickly, so the company stays one step ahead of competitors who can copy features within days.
“there are no product advantages that are longlasting anymore”
Staying one step ahead only becomes a durable advantage if customers expand their use cases faster and more deeply with the vendor. Listen
Cassie says businesses that stay a couple steps ahead win by getting customers to expand use cases faster and more deeply than anywhere else. The customer information on the platform fuels reinforcement learning and raises switching costs. She says the proof is whether customers follow and spend more over time, and if that is not visible, it is a warning sign.
“the businesses that do this really well, they stay a couple steps ahead in a manner that wants customers to expand the use cases faster and more deeply than they will anywhere else”
Listen to the episode Strategy & market Link to this Report a problem
Implementation is moving out of customer success and into the product org in some portfolio companies. Listen
Cassie says she is seeing a notable organizational change in portfolio companies: implementation, previously in the chief customer officer or post-sales lane, is moving into product. Companies view getting customers up and running as a product opportunity, and the best organizations productize what they learn from forward-deployed work immediately.
“implementation in many places is being moved out of the chief customer officer lane or the post-sales lane and into product”
Customer success has promised to be proactive rather than reactive since the dawn of software, and Cassie says that is not happening yet. Listen
Cassie says customers fall out of sync with configuration when they replatform, causing repeated breaks and health checks for customers who have been with the company for four years. She says proactive customer success only becomes real when an agentic product spots configuration problems before they come up. She says this is where her energy is.
“customer success has promised to be proactive versus reactive since the dawn of software. It's not happening right”
Listen to the episode Retention & customer success Link to this Report a problem
Bigger venture funds tend to underperform mainly because partners end up doing more deals and firms drift into later stages. Listen
Cassie says the data shows larger funds return less, and that the usual causes are that each partner deploys more capital, so instead of two or three deals a year they may do six, and that firms suffer strategy creep into Series A. Primary's response is to scale by adding check writers so no single partner does more deals per fund. This describes how VC firms are built rather than something a founder can act on.
“academically speaking the bigger the funds the worse the returns”
Listen to the episode Fundraising & investors Link to this Report a problem